What problem does it solve? Building a reliable consumer for Alpaca Broker API real-time event streams is error-prone: connections silently die, events are missed during downtime, and duplicate deliveries corrupt local state. This Skill provides the connection, replay-cursor, reconnection, and idempotency patterns needed to process account, journal, funding, trade, and non-trade-activity events without data loss. ## Core Features & Use Cases - Stream coverage: Documents all five SSE streams (account status, journals, funding/transfers, trades, non-trade activities) including their inconsistent v1/v2 paths and ULID versus integer event ID cursors. - Replay and recovery: Explains since/since_id/since_ulid cursors so a reconnecting consumer replays missed events instead of silently dropping them. - Reliability patterns: Covers heartbeat/silence detection, exponential backoff with a single-reconnect guard, and an idempotent processing pipeline keyed on event_id. - Use Case: You are building a backend that must update local order and transfer state the moment Alpaca fills an order or executes a journal. Use this Skill to implement an SSE consumer that survives deploys and network drops, deduplicates replays, and pairs with periodic reconciliation. ## Quick Start Use the alpaca-broker-sse-events skill to implement a resilient SSE consumer for Alpaca trade update events with reconnect and idempotent processing.